In the past decade, several CNS-based neural prosthetics have achieved remarkable clinical outcomes. For example, deep brain stimulation (DBS) is now routinely used to treat Parkinsonian tremors and cochlear prosthetics stimulate the auditory nerve to restore high levels of hearing to the profoundly deaf. However, not all devices have achieved the same level of success. For example, retinal prosthetics do not reliably elicit complex (or even simple) spatial percepts even though individual electrodes can consistently elicit focal percepts (phosphenes). Even the more successful applications can sometimes be inconsistent and have created unwanted side effects. To improve the outcomes associated with retinal (and other neural) prosthetics, we are studying the fundamental interactions between electric stimulation and targeted neurons using a combination of electrophysiology, anatomy and computational modeling. In retinal neurons, we found that the lowest thresholds in response to electric stimulation occur when the stimulating electrode is positioned over the proximal portion of the axon (near the soma). This region of low threshold is precise aligned with a dense band of voltage-gated sodium channels identified immunochemically. Different types of retinal neurons have different bands (lengths and locations) as well as different absolute (lowest) thresholds. This implies that band differences contribute to threshold differences and we propose to investigate the details by which this occurs. It is likely that the band is also the site in which spikes are initiated but this has not yet been confirmed;here, we propose several experiments to determine whether the band is in fact the site. Knowledge of the site of spike initiation enables us to systematically study how changes to the applied electric field (arising from the stimulus pulse) alter the neural response. Our preliminary data suggests that analogous to stimulation of axons, the second spatial derivative of the induced voltage profile along the band is a good predictor of pulse efficacy. In the temporal domain, we want to determine which stimulus profile is most effective for activating different elements within the neuron. We are also exploring whether specific classes or sub-classes of neurons respond preferentially to different stimulus frequencies. Our hope is that by understanding the basic elements of the response mechanism, including the cause of response differences between different types of neurons, we can develop more effective stimulation methods that will lead to better clinical outcomes. PUBLIC HEALTH RELEVANCE: Despite remarkable progress with some CNS-based neural prosthetics, devices that target other regions of the CNS have not yet achieved satisfactory clinical results. As a step towards developing more effective stimulation methods for retinal prosthetics, we are studying the fundamental interactions between electric stimulation and targeted retinal neurons. Because the knowledge acquired in these studies pertains to general activation mechanisms, it is likely to be applicable to both the retinal prosthetic as well as a wide range of other CNS- based devices.